Valerie LeMay

The roles of nearest neighbor methods in imputing missing data in forest inventory and monitoring databases.

Abstract Almost universally, forest inventory and monitoring databases are incomplete, ranging from missing data for only a few records and a few variables, common for small land areas, to missing data for many observations and many variables, common for large land areas. For a wide variety of applications, nearest neighbor (NN) imputation methods have been developed to fill in observations of variables that are missing on some records (Y-variables), using related variables that are available for all records (X-variables). This review attempts to summarize the advantages and weaknesses of NN imputation methods and to give an overview of the NN approaches that have most commonly been used. It also discusses some of the challenges of NN imputation methods. The inclusion of NN imputation methods into standard software packages and the use of consistent notation may improve further development of NN imputation methods. Using X-variables from different data sources provides promising results, but raises the issue of spatial and temporal registration errors. Quantitative measures of the contribution of individual X-variables to the accuracy of imputing the Y-variables are needed. In addition, further research is warranted to verify statistical properties, modify methods to improve statistical properties, and provide variance estimators.

Imputing tree-lists from aerial attributes for complex stands of south-eastern British Columbia

The nearest neighbor, k-nearest neighbors, distance-weighted k-nearest neighbor, and class-weighted k-nearest neighbor imputation methods were compared for accuracy in estimating tree-lists (list of species and diameter for each tree) from aerial attributes for complex stands, with up to nine species and a wide range of sizes, in south-eastern British Columbia, Canada. For the four imputation methods, the most similar neighbor distance metric was used, and three neighbors were used for the k-nearest neighbor methods. Ground variables used to represent the tree-list included the number of trees per hectare by species, ranges of diameter by species, and basal area per hectare. Aerial variables included species composition, crown closure (%), elevation, biogeoclimatic ecosystem classification (BEC) zones, height, age, and site class. Sample data were divided, and the imputation methods were compared for accuracy using observed and estimated species composition, stand tables, basal area, and volume per hectare. Also, the imputed tree-list was used to predict yield using a stand level growth model, and this predicted yield was compared to the yield obtained using the actual tree-list. Of the four approaches used, the nearest neighbor was marginally better, but the methods that averaged the three nearest neighbors were somewhat better for the distribution of stems per hectare by diameter for the more sparse hardwood species. Of the three averaging methods, weighting by similarity of the species composition and the BEC zone provided better results. In using the estimated trees lists in a growth and yield model, the average volumes were reasonable at the beginning and end of the period for all methods. However, the volumes for a particular stand could be quite different than that obtained for an observed tree-list.

Modelling and prediction of dominant height and site index of Eucalyptus globulus plantations using a nonlinear mixed-effects model approach

A nonlinear mixed-effects model approach was used to model dominant height and site index for Eucalyptus globulus Labill. plantations in southeastern Australia. Mixed effects were considered initially for all three parameters of a modified Chapman–Richards model. Inclusion of random effects improved fitting and accounted for the within-plot heteroscedasticity. To correct for within-plot autocorrelation, a power autocorrelation model allowing for irregular intervals for remeasurements was found to be most appropriate. Additional fertilizer application at age 1 year and a number of environmental variables were related to the fixed-effects parameters, but these were not statistically significant, whereas mean annual rainfall and average daily maximum temperature in July (winter) greatly reduced the residual variability among plots. The resulting nonlinear mixed-effects model combines dominant height and site index prediction into a single model and predicts polymorphic height growth rates on different sites....

Prediction Models for Estimating the Area, Volume, and Age of Rubber (Hevea brasiliensis) Plantations in Malaysia using Landsat TM Data

SUMMARY Rubber tree (Hevea brasiliensis (Wild. ex Adr. de Juss.) Muell Arg.) plantations in Malaysia are important sources of natural rubber and wood products. Effective management and appropriate policy for these resources requires reliable forecasts of resource availability. However, to achieve these goals, effective inventories are required. This promoted research into supplementing ground-based survey methods with satellite remote sensing information. A study was conducted to investigate the relationship between Landsat Thematic Mapper (TM) data and rubber stand parameters and to develop and evaluate models for estimating area, volume, and age of rubber plantations. Statistically significant models for estimating volume and age of rubber stands were obtained. For volume models, the R2 values were all higher than 0.70 and standard error of the estimate (SEE) values were lower than 54 m3/ha. R2 and SEE values achieved from age models evaluated ranged from 0.34–0.64 and 6.4–8.2 years. A logistic regression model produced classifications with an accuracy of 87% for predicting the presence of rubber plantations. Thus, Landsat TM provides an acceptable data source for estimating wood volume and stand age, and for predicting the presence of rubber plantations.

Modeling lodgepole pine radial growth relative to climate and genetics using universal growth‐trend response functions

Forests strongly affect Earths carbon cycles, making our ability to forecast forest-productivity changes associated with rising temperatures and changes in precipitation increasingly critical. In this study, we model the influence of climate on annual radial growth using lodgepole pine (Pinus contorta) trees grown for 34 years in a large provenance experiment in western Canada. We use a random-coefficient modeling approach to build universal growth-trend response functions that simultaneously incorporate the impacts of different provenance and site climates on radial growth trends under present and future annual (growth-year), summer, and winter climate regimes. This approach provides new depth to traditional quantitative genetics population response functions by illustrating potential changes in population dominance over time, as well as indicating the age and size at which annual growth begins declining for any population growing in any location under any present or future climate scenario within reason, given the ages and climatic conditions sampled. Our models indicate that lodgepole pine radial-growth levels maximize between 3.9 degrees and 5.1 degrees C mean growth-year temperature. This translates to productivity declining by the mid-21st century in southern and central British Columbia (BC), while increasing beyond the 2080s in northern BC and Yukon, as temperatures rise. Relative to summer climate indices, productivity is predicted to decline continuously through the 2080s in all locations, while relative to winter climate variables, the opposite trend occurs, with the growth increases caused by warmer winters potentially offsetting the summer losses. Trees from warmer provenances, i.e., from the center of the species range, perform best in nearly all of our present and future climate-scenario models. We recommend that similar models be used to analyze population growth trends relative to annual and intra-annual climate in other large-scale provenance trials worldwide. An open-access growth-trend data set encompassing numerous biomes, species, and provenances would contribute substantially to predicting forest productivity under future-climate scenarios.

Vertical partitioning between sister species of Rhizopogon fungi on mesic and xeric sites in an interior Douglas‐fir forest

Understanding ectomycorrhizal fungal (EMF) community structure is limited by a lack of taxonomic resolution and autecological information. Rhizopogon vesiculosus and Rhizopogon vinicolor (Basidiomycota) are morphologically and genetically related species. They are dominant members of interior Douglas‐fir (Pseudotsuga menziesii var. glauca) EMF communities, but mechanisms leading to their coexistence are unknown. We investigated the microsite associations and foraging strategy of individual R. vesiculosus and R. vinicolor genets. Mycelia spatial patterns, pervasiveness and root colonization patterns of fungal genets were compared between Rhizopogon species and between xeric and mesic soil moisture regimes. Rhizopogon spp. mycelia were systematically excavated from the soil and identified using microsatellite DNA markers. Rhizopogon vesiculosus mycelia occurred at greater depth, were more spatially pervasive, and colonized more tree roots than R. vinicolor mycelia. Both species were frequently encountered in organic layers and between the interface of organic and mineral horizons. They were particularly abundant within microsites associated with soil moisture retention. The occurrence of R. vesiculosus shifted in the presence of R. vinicolor towards mineral soil horizons, where R. vinicolor was mostly absent. This suggests that competition and foraging strategy may contribute towards the vertical partitioning observed between these species. Rhizopogon vesiculosus and R. vinicolor mycelia systems occurred at greater mean depths and were more pervasive in mesic plots compared with xeric plots. The spatial continuity and number of trees colonized by genets of each species did not significantly differ between soil moisture regimes.

Elliptical Time-Density model to estimate wildlife utilization distributions

Summary 1. We present a new animal space-use model (elliptical time density – ETD) that uses discrete-time tracking data collected in wildlife movement studies. The ETD model provides a trajectory-based, nonparametric approach to estimate the utilization distribution (UD) of an animal, using model parameters derived directly from the movement behaviour of the species. The model builds on the theory of ‘time-geography’ whereby elliptical constraining regions are established between temporally adjacent recorded locations. 2. Using a Weibull speed distribution fitted for an animal’s movement data, a time density value (i.e. time per unit landscape) is determined from the expectation of all elliptical regions equal to, or greater than, the minimum bounding ellipse for a given landscape point. We tested the ETD model using a tracking dataset for an African elephant (Loxodonta africana) and compared the resulting UDs for regularly sampled, frequently recorded locations, as well as irregular random time intervals between locations and also infrequent temporal-sampling regimes, providing insight to the method’s performance with different resolution data. We compared the performance of the ETD model, the Brownian bridge movement model (BBMM), the time-geography density estimator (TGDE) and the Kernel Density Estimator (KDE) by calculating omission/commission errors from the predicted space-use distribution of each model relative to the true known UD of our elephant test data. The comparison was made for the 10–99% percentile UD model areas. 3. The ETD90 model (i.e. ETD model parameterized using the 90% percentile value of the Weibull speed distribution) resulted in the fewest errors of commission and omission with regard to locating the true movement path at the 99% percentile UD area. 4. The ETD model provides an improved approach for estimating animal UDs since (i) parameters are derived directly from the tracking data rather than assumed; (ii) parameter values are biologically interpretable; (iii) the Weibull speed distribution is adaptable to various temporal-sampling regimes; and (iv) the ETD model handles the case of degenerate ellipses thus preserving landscape connectivity in the UD. Software (freeware) for calculating the ETD and a Bayesian framework for estimating the Weibull distribution speed parameters are also introduced in the paper.

Development and testing of diameter increment models for mixed swamp forests of Sarawak

Abstract The variability of diameter increment of commercial species of the mixed swamp forests of Sarawak is high, partly because of the large number of species represented. It may be desirable to combine several species together to obtain a diameter-increment estimation equation because of the cost of sampling and the fact that some species may occur infrequently, and for simplicity. In this study, a diameter-increment model was developed and fitted for three different modelling levels, defined as individual species, two groups of species (light-demanding versus shade-tolerant species) and one group of all species combined. These three modelling levels were compared using estimated standard error, Pseudo R 2 , mean prediction error, mean-squared prediction error and mean absolute prediction error based on validation data. Modelling at the individual-species level resulted in only slightly better predictions compared with the one-group level; no real gain was noted for the two-group level compared with the one-group level. For most species, the one-group level resulted in predictions similar to the individual-species models. The one-group level is recommended, although some individual-species models are useful for those species which give poor predictions using the one-group model.

The Relative Abundance of Mountain Pine Beetle Fungal Associates Through the Beetle Life Cycle in Pine Trees

The mountain pine beetle (MPB) is a native bark beetle of western North America that attacks pine tree species, particularly lodgepole pine. It is closely associated with the ophiostomatoid ascomycetes Grosmannia clavigera, Leptographium longiclavatum, Ophiostoma montium, and Ceratocystiopsis sp.1, with which it is symbiotically associated. To develop a better understanding of interactions between beetles, fungi, and host trees, we used target-specific DNA primers with qPCR to assess the changes in fungal associate abundance over the stages of the MPB life cycle that occur in galleries under the bark of pine trees. Multivariate analysis of covariance identified statistically significant changes in the relative abundance of the fungi over the life cycle of the MPB. Univariate analysis of covariance identified a statistically significant increase in the abundance of Ceratocystiopsis sp.1 through the beetle life cycle, and pair-wise analysis showed that this increase occurs after the larval stage. In contrast, the abundance of O. montium and Leptographium species (G. clavigera, L. longiclavatum) did not change significantly through the MPB life cycle. From these results, the only fungus showing a significant increase in relative abundance has not been formally described and has been largely ignored by other MPB studies. Although our results were from only one site, in previous studies we have shown that the fungi described were all present in at least ten sites in British Columbia. We suggest that the role of Ceratocystiopsis sp.1 in the MPB system should be explored, particularly its potential as a source of nutrients for teneral adults.

Estimating Carbon Stocks and Stock Changes in Forests: Linking Models and Data Across Scales

The increasing amount of atmospheric carbon has been linked with changes in climate, prompting efforts to reduce the amount of carbon emissions. Estimates of forest carbon stocks and stock changes are needed, along with how these change over time, and how sequestration might be increased through forest management activities such as afforestation, reforestation, stand management, and forest protection. Carbon is accrued through increased live biomass and/or increased dead organic matter and soil carbon, whereas carbon is released to the atmosphere through respiration, decomposition, and burning. For large land areas, estimating the amount of carbon sequestration into and out of a forest system involves integrating a number of data sources and models at a variety of spatial and temporal scales. The methods used to integrate data and models across time and spatial scales vary. In this paper, we present a discussion of methods used to obtain information on carbon stocks for very large land areas, using reported analyses as examples.